© 2004 Nature Publishing Group http://www.nature.com/nsmb
ARTICLES
A mutant chaperonin with rearranged inter-ring
electrostatic contacts and temperature-sensitive
dissociation
B Trevor Sewell1, Robert B Best1,4, Shaoxia Chen2,4, Alan M Roseman2,4, George W Farr3, Arthur L Horwich3 &
Helen R Saibil2
The chaperonin GroEL assists protein folding through ATP-dependent, cooperative movements that alternately create folding
chambers in its two rings. The substitution E461K at the interface between these two rings causes temperature-sensitive,
defective protein folding in Escherichia coli. To understand the molecular defect, we have examined the mutant chaperonin by
cryo-EM. The normal out-of-register alignment of contacts between subunits of opposing wild-type rings is changed in E461K to
an in-register one. This is associated with loss of cooperativity in ATP binding and hydrolysis. Consistent with the loss of negative
cooperativity between rings, the cochaperonin GroES binds simultaneously to both E461K rings. These GroES-bound structures
were unstable at higher temperature, dissociating into complexes of single E461K rings associated with GroES. Lacking the
allosteric signal from the opposite ring, these complexes cannot release their GroES and become trapped, dead-end states.
The chaperonin GroEL is an 800-kDa, homo-oligomeric double-ring
assembly that provides ATP-dependent kinetic assistance to protein
folding1–4. It binds non-native polypeptide in an open ring via exposed
hydrophobic surfaces that are contacted by hydrophobic side chains
lining the GroEL central cavity. Then, upon binding of the lid-like
cochaperonin ring, GroES, to the same ring as polypeptide (cis), the
substrate protein is released from the cavity wall into the nowencapsulated and hydrophilic central cavity where productive folding
occurs. ATP hydrolysis in the folding-active ring is followed by ATP
binding in the opposite (trans) ring, which allosterically triggers ejection of the cis ligands: GroES, polypeptide and ADP5. For any given
cycle of the reaction, released polypeptides that have reached native
form or a committed state (no longer recognizable by GroEL) proceed
to assemble, if required, and carry out their biological functions.
Polypeptides that do not reach native form can be rebound by GroEL to
further attempt correct folding or, alternatively, can bind to other
chaperones or proteases in the same cellular compartment in a kinetic
partitioning mechanism.
As indicated, ATP binding and hydrolysis drive the chaperonin reaction cycle. ATP binding occurs in a cooperative manner: there is positive cooperativity within a ring of seven subunits6, but negative
cooperativity between the rings, such that ATP binds tightly to only
one ring at a time7. Because GroES associates only with an ATP-bound
ring, asymmetric GroEL–GroES complexes are populated, with a
minority of GroES–GroEL–GroES complexes7–9. Notably, as with
many other molecular machines, the binding of ATP carries out the
major work of this machine: in cis, binding of ATP and GroES triggers
productive folding, and in trans, the binding of ATP sends an allosteric
signal that ejects the cis ligands. ATP hydrolysis, by contrast, is used to
advance the machine forward through its reaction cycle, weakening
the affinity of GroEL for GroES on the cis ring and ‘priming’ the transtriggered release of GroES. At the same time, ATP hydrolysis in the
cis ring allosterically adjusts the trans ring to accept non-native
polypeptide—that is, the cis-ATP GroEL–GroES complex lacks affinity
for polypeptide, but after cis hydrolysis polypeptide is accepted into
the open trans ring of a cis-ADP complex.
Both the negative cooperativity between rings with respect to ATP
binding and the allosteric signaling from an ATP-bound trans ring that
leads to ejection of the cis ligands depend on inter-ring contacts for
proper signaling10. An additional long-range signal sent across the
ring-ring interface produces the switch of affinity for polypeptide of
the trans ring that occurs after cis ATP hydrolysis. Finally, the simultaneous binding of both ATP and polypeptide to an open trans ring has
been observed to accelerate the ejection of cis ligands5. This indicates
that not only adenine nucleotide but also non-native polypeptide can
send an allosteric signal across the ring-ring interface. All of these critical inter-ring signaling events in GroEL are transmitted through two
equatorially projecting portions of each subunit that make contacts
1Electron Microscope Unit and Department of Chemistry, University of Cape Town, Rondebosch, South Africa. 2Crystallography Department, Birkbeck College,
London WC1E 7HX UK. 3Department of Genetics and Howard Hughes Medical Institute, Yale University School of Medicine, New Haven, Connecticut 06510 USA.
4Present addresses: Laboratory of Chemical Physics, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda,
Maryland 20892-0520, USA (R.B.B.) and Medical Research Council Laboratory of Molecular Biology, MRC Centre, Cambridge CB2 1QH, UK (S.C. and A.M.R.).
Correspondence should be addressed to H.R.S. (h.saibil@mail.cryst.bbk.ac.uk).
Published online 10 October 2004; doi:10.1038/nsmb844
1128
VOLUME 11 NUMBER 11 NOVEMBER 2004 NATURE STRUCTURAL & MOLECULAR BIOLOGY
© 2004 Nature Publishing Group http://www.nature.com/nsmb
ARTICLES
Figure 1 Structural comparison of wild-type
(WT) and E461K GroEL. (a) Cryo-EM map of wildtype GroEL9. Three subunits forming a set of
contacts across the ring interface are outlined.
(b) Equivalent map of GroEL E461K. In this
case the inter-ring contacts are between pairs
of subunits and two are outlined. (c) Docking
of wild-type rings that have been rotated to the
energy-minimized position shows that there are
no large changes in the structure except the rigidbody rotation of one ring relative to the other.
(d–f) Slabs of the atomic structures centered at
the interface, highlighting the residues involved
in salt bridge contacts. The boxed area in a is
enlarged in d. (d) The inter-ring interface of wildtype GroEL20. The left-hand salt bridge contains
Glu434 and Lys105, and the right-hand one
contains Arg452 and Glu461. (e) The interface
of E461K, modeled by energy minimization.
An alternative set of salt bridges is predicted,
primarily between Glu434 and Lys470 (left and
right contacts) and Glu102 and Lys105 (only
one pair is formed because of steric hindrance
at the central contact). (f) The interface between
thermosome α subunits (PDB entry 1A6D12) for
comparison. Panels c–f were created with PyMOL
(http://www.pymol.org).
chaperonin complex and in enabling proper
cooperative signaling.
with two adjacent subunits of the opposite ring, with a 1:2 subunit
relationship (Fig. 1a,d). The nature of the allosteric transmission
through these sites is suggested by cryo-EM analysis of the
GroEL–ATP complex11. In that study, ATP binding was shown to distort the ring interface with an overall increase in separation of the
rings and tilts of the equatorial domains that change the relative distances of the inter-ring contacts. These distortions are not observed in
the crystal lattices of various GroEL–nucleotide complexes.
To understand the structural and mechanistic basis of GroEL
cooperativity, which is central to its function in assisting protein
folding, we examined a GroEL mutant with defective cooperativity.
Here we report a structural analysis of GroEL E461K showing a
marked reorganization of the inter-ring interface, causing it to
resemble the group II, archaeal and eukaryotic chaperonin
structure, in which the interface is formed by a 1:1 subunit interaction12. In vitro studies reveal complete loss of cooperativity of this
assembly with respect to ATP binding and hydrolysis. In the presence of ATP and GroES we observed efficient formation of football
structures, with GroES bound at either end of the 461 cylinder,
reflecting the loss of negative cooperativity. Notably, at elevated
temperature these complexes dissociated into complexes of single
GroEL rings in complex with GroES, a dead-end state that cannot
release its ligands. These observations reveal the importance of
proper ring-ring alignment both in maintaining the stability of the
RESULTS
An in vivo genetic screen had isolated a temperature-sensitive point
mutant of GroEL, in which Glu461 in the equatorial interface region
was altered to lysine13. The E461K mutation alters an electrostatic
contact formed with Arg452 to a repulsion and was associated
in vivo with reduced stability of the assembly, observed as ‘smearing’
of the GroEL-immunoreactive pattern on western blots of native
gels, and with temperature-sensitive behavior in several protein
folding assays. Defects observed in subsequent in vitro studies in
polypeptide release were notable in view of the remote position of
residue 461 from the apical binding sites some 50–100 Å from the
ring interface14.
To study the structure of the mutant complex, we overexpressed it in
E. coli, purified it, and examined it by cryo-EM and image reconstruction. We also examined its activities in vitro in ATP turnover and protein folding. In addition, the proportions of different assemblies
Figure 2 Calculation of the energy of the ring interface in wild-type and
mutant GroEL as a function of angle of rotation relative to the position found
in wild-type GroEL. The plot covers the asymmetric unit, one-seventh of a
complete turn, shown from –26° to +26° of rotation. The energy minimum
for the wild-type protein is close to 0°, corresponding to the ring location in
wild-type GroEL. For the E461K mutant it is at 16°, corresponding to the
rings being in register, consistent with the structure in Figure 1b.
NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 11 NUMBER 11 NOVEMBER 2004
1129
ARTICLES
© 2004 Nature Publishing Group http://www.nature.com/nsmb
Figure 3 Initial rate of ATPase activity as a function of ATP concentration.
The sigmoidal shape of the curve for wild-type (WT) GroEL (inset) results
from positive cooperativity of ATP binding, and the reduced ATPase at
higher concentrations reflects the negative cooperativity between the GroEL
rings. Both these effects are absent in the mutant. The continuous curves
were obtained by fitting the data to the Hill equation. In the case of the
wild type, the fit is an approximation that does not take into account the
effect of negative cooperativity on the shape of the curve at higher ATP
concentrations. At the KCl concentration used (5 mM), this effect is small.
formed with GroES at permissive and heat shock temperatures were
assessed by negative-stain EM.
GroEL E461K has a rearranged inter-ring interface
The cryo-EM structure of GroEL E461K at a resolution of 24.5 Å
clearly reveals the main structural consequence of this mutation.
Although the maps of wild-type and E461K GroEL are superficially
similar (Fig. 1a,b), examination of the relationship between the
rings reveals a complete reorganization of the ring interface in the
mutant. The opposing subunits of E461K rings are in register, and
the contacts are 1:1 instead of 1:2 (Fig. 1b). At this resolution the
maps appear identical except for an 18° rigid body rotation about
the seven-fold symmetry axis of one ring relative to the other in the
mutant (Fig. 1c). The contact regions between rings are enlarged
and shown as atomic structures in Fig. 1d–f. The E461K interface
(Fig. 1e) is derived from an energy-minimized model (see below),
and the group II thermosome interface is shown for comparison
(Fig. 1f).
Computation of the energy as a function of the angle of rotation
about the seven-fold symmetry axis for wild-type and mutant
GroEL gives results that are consistent with the EM reconstructions
(Fig. 2). There is an energy minimum close to 0° for wild-type
GroEL, corresponding to the offset ring conformation of the normal
structure (Fig. 1a,d). For E461K, a minimum is seen at 16°, corresponding to the in-register alignment seen in the mutant structure
(Fig. 1b,c). The electrostatic contributions to the overall energy were
dominant. The energy-minimized model of the E461K interface
enables identification of ion pair residues likely to interact in the
mutant (Fig. 1e). The model predicts that Lys470 and Glu102,
residues near the interface but not involved in any contacts in the
wild-type structure, are recruited to form contacts, and that the
mutant Lys461 side chain rotates away from the contact region. The
predicted salt bridges are formed primarily between Glu434 and
Lys470, and between Glu102 and Lys105. Other residues, including
Asp435, Arg445 and Arg452, are close enough to make a substantial
contribution to the interaction.
Defective activity in vitro
ATPase activity is increased in the E461K mutant relative to wild-type
GroEL across a range of ATP concentrations (Fig. 3), reflecting the loss
of negative cooperativity. The nonsigmoidal shape of the ATPase curve
at low ATP concentration (Fig. 3, inset) indicates that positive cooperativity is also lost in the mutant. Chaperonin function was tested by
assaying the refolding of denatured mitochondrial MDH. At 25 °C, the
mutant chaperonin supported MDH refolding, but this was defective
Figure 4 Temperature sensitivity of GroEL E461K–GroES complexes in
the presence of 4 mM ATP observed by negative-stain EM. (a) Complexes
incubated at 23 °C. (b) Complexes incubated at 40 °C. (c) Complexes
incubated at 40 °C, then cooled down to 4 °C. The cartoon symbols indicate
the predominant complexes present in each sample. In a, side views of
GroEL, GroEL–GroES (bullet-shaped complexes) and GroES–GroEL–GroES
(football-shaped complexes) are visible, as well as end views. The complexes
in b are mostly end views. These are attributed to football complexes that
have split into single GroEL rings, each with a bound GroES (see text).
Upon cooling (c), the split complexes reassociate and mainly football
complexes are seen.
1130
VOLUME 11 NUMBER 11 NOVEMBER 2004 NATURE STRUCTURAL & MOLECULAR BIOLOGY
© 2004 Nature Publishing Group http://www.nature.com/nsmb
ARTICLES
Figure 5 Allosteric communication in wild-type (WT) and E461K GroEL. (a) Schematic diagram of three equatorial domains (green) at the front of the GroEL
complex, showing the staggered contacts (red and blue boxes) made across the ring interface. Helix D (residues 87–109) is shown as a coil. The view is as
in Figure 1d. On the right, the domain tilting accompanying ATP binding in one ring11 is represented. An additional domain in the upper ring is included to
show the conformational strain between adjacent equatorial domains within the ATP-bound ring. Routes of communication of conformational changes are
indicated by purple dashed arrows. Both the intra- and inter-ring interfaces between adjacent equatorial domains are distorted by the domain tilting. The
arrow extending from the ATP-binding pocket represents the propagation of conformational changes to the intermediate and apical domains. (b) The
situation in GroEL E461K is simpler, with the equatorial domains in an eclipsed conformation at the interface. Three salt-bridge contacts are predicted by
the simulation (Figs. 1e and 2). Helix D connects to the central contact. Owing to the lack of negative cooperativity, ATP fills both rings simultaneously. The
ring interface weakens, but changes communicated to the central contact are unlikely to lead to domain tilting.
at 37 °C, with no recovery of MDH activity at 1 mM ATP, and ∼50%
recovery at 10 mM ATP (data not shown). Similar results have been
obtained previously15, although the loss of positive cooperativity was
not resolved in that study.
Structural effects of ATP and elevated temperature
Incubation at temperatures up to 60 °C does not cause any changes in
the structures of isolated wild-type or E461K GroEL observable by
negative-stain EM (data not shown). However, substantial effects were
observed when we compared the proportions of different assemblies in
preparations of wild-type and E461K GroEL incubated with GroES and
ATP, at 23 and 40 °C. The loss of negative cooperativity in E461K is
reflected in the higher proportion of double-sided, GroES–
GroEL–GroES ‘football’ complexes of the mutant protein (Fig. 4a) than
in the corresponding wild-type preparation. For the mutant at 40 °C,
the football views gradually disappear with increasing incubation time,
and are replaced by relatively featureless rings, many lacking the characteristic seven-fold symmetry of GroEL end views, and also by triangular
views (Fig. 4b). The footballs rapidly reappear upon cooling (Fig. 4c).
These findings were verified by counting large data sets in a systematic
time course experiment on wild-type and E461K complexes incubated
at 23 and 40 °C, and then cooled to 4 °C. The counts show that the proportion of football plus bullet side views is in the range expected for wild
type (data not shown) at 40 °C (30%), low in E461K at 23 °C (4%) and
very low in E461K at 40 °C (1.5%), but that it increases markedly upon
cooling of the E461K samples to 4 °C (15%). The heated mutant sample
contains a great majority of end views (Fig. 4b).
The featureless rings and triangular views can be identified as complexes of GroES with a single ring of GroEL. These ‘half-footballs’ are
fairly round particles, which adopt a more uniform distribution of orientations on the EM grid than the cylindrical complexes containing
double-ring GroEL. The triangles can be recognized as side views of
the cis cavity of GroEL–GroES complexes, or half-footballs. Although
images of double-ring complexes in the presence of ATP or of ATP and
GroES were scarce, we obtained cryo-EM reconstructions of these
complexes. Consistent with a disordered interface between the two
rings, these maps were of poor quality, with rotationally smeared features (data not shown).
DISCUSSION
In this study we have described a GroEL chaperonin mutant in which
the normal inter-ring electrostatic contacts have been replaced, as the
result of mutation of a charge attraction to one of repulsion, with a new
set of electrostatic contacts that position the two rings 18° out of normal register. The price paid for such a rearrangement is complete loss
of cooperativity with respect to ATP binding and hydrolysis. Both the
negative cooperativity between rings and also the positive cooperativity within them are abolished, supporting the idea that the lines of
cooperative communication between rings are inextricably linked to
communication between subunits within the rings16. Despite the complete loss of cooperativity with respect to ATP, this assembly remains
partly functional in protein folding, observed both in vivo during the
earlier study of the temperature-sensitive mutant13 and here and elsewhere15 in vitro. That is, there is sufficient residual allosteric function
to allow productive cycles of binding and release of substrate polypeptide. How does the temperature sensitivity come about?
Considering the formation of the folding chamber, we know from
studies of single-ring GroEL (SR1) that binding of non-native substrate,
followed by ATP-and GroES-induced conformational change, can lead
to the normal folding of substrate inside the cavity associated with one
round of ATP hydrolysis5,17. The result of these steps is a stable, deadend complex, SR1–ADP–GroES, with the folded substrate trapped
inside the cavity. It seems likely that the steps resulting in formation of
the GroEL–GroES chamber occur fairly normally in GroEL E461K.
However, owing to the loss of negative cooperativity with respect to ATP,
GroES binds simultaneously to both rings of the E461K assembly, so
that football complexes predominate. The football conformation is
likely to place extra strain on the equatorial domains18.
The problem in E461K arises in the next phase of the functional
cycle. In wild-type complexes, ATP hydrolysis in the GroES-bound cis
ring weakens the binding of GroES but does not result in its release5.
Subsequently, ATP binding to the trans ring triggers the release of
GroES, allowing the escape of folded substrate and trapped ADP. This
communication from trans to cis rings must occur through conformational changes in the ring interface. In wild-type complexes, ATP binding weakens the ring interface and distorts it by tilting the equatorial
domains (ref. 11; N.A. Ranson (Biochemistry and Microbiology
NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 11 NUMBER 11 NOVEMBER 2004
1131
© 2004 Nature Publishing Group http://www.nature.com/nsmb
ARTICLES
Department, Leeds University), D.K. Clare (Crystallography
Department, Birkbeck College), G.W.F., A.L.H. and H.R.S, unpublished data). In E461K complexes, however, the ring interface is less
able to support the distorted and weakened interactions of nucleotidebound states, and thus the ATP- and GroES-bound complexes are
much less stable than in the case of the wild-type protein. Indeed, at
elevated temperature, the football complexes fall apart into single-ring
GroEL–GroES complexes (Fig. 3b). Without the apposed equatorial
domains of an opposite ring to provide the trigger, GroES release from
these single-ring complexes is greatly impaired, as in the designed SR1
complex. The dissociation of the interface is reversed upon cooling,
with the rapid reappearance of footballs (Fig. 3c). Notably, this
marginally stable, wrongly arranged inter-ring interface can support
some degree of normal cycling at reduced temperature.
Insight into the ability of the mutant to retain some normal GroEL
function despite the complete rearrangement of the ring interface
contacts and its instability comes from an unexpected parallel with a
different subfamily of chaperonins. Notably, the 1:1 arrangement of
subunits across the interface has a natural precedent. An interface
with in-register subunit contacts occurs in the group II chaperonins,
such as the archaeal thermosome12. Because the group II chaperonins
also exhibit negative cooperativity19, it must be transmitted across the
ring interface by different interactions than the ones in E461K. The
mutant interface bears a marked resemblance to that of the thermosome (Fig. 1f), except that the inter-ring separation in the group II
structure is smaller, and the interaction seems tighter. Perhaps the
stronger interface enables this chaperonin subfamily to withstand
nucleotide-induced distortions of the equatorial contacts in the 1:1
geometry without excessive weakening of the contacts.
These findings provide new insights into the mechanism of negative
cooperativity in chaperonins. From our earlier work on GroEL–ATP
complexes11, we know that the equatorial domains tilt upon ATP
binding, and that the interface between the rings is distorted. The
ATP-binding site is linked through helix D (residues 87–109) to the left
inter-ring contact (Figs. 1d and 5a). In the E461K mutant, helix D
does not connect to an inter-ring contact at the side of the domain
(Figs. 1e and 5b) and there is no negative cooperativity. This comparison suggests that negative cooperativity depends on equatorial
domain tilting and interface distortion transmitted through helix D to
one side of the domain, providing leverage for tilting. We predict that
the equatorial domains in group II chaperonins will also tilt when ATP
is bound in one ring, because the ATP binding site is directly linked to
the left and right contacts (Fig. 1f).
The positive and negative cooperativity mechanisms are clearly
interrelated in the central role of the nucleotide-binding, equatorial
domains, which form the major intra-ring contacts and all the interring contacts20. By virtue of its unanticipated effects on the assembly
of the complex, the E461K mutation has provided us with a novel and
informative probe of the levers operating the intricate allosteric mechanism of chaperonins.
METHODS
Protein purification and activity assays. E461KGroEL E461K was overexpressed and purified as described19. The buffer for EM contained 20 mM Tris,
5 mM KCl, 10 mM MgCl2, pH 7.4. ATPase activity was measured as
described21. Refolding of MDH was carried out as described22.
Cryo-EM. Specimens containing 1 µM GroEL E461K were flash-frozen in liquid ethane and imaged at –170 °C on a JEOL 2010 electron microscope at
200 kV using an Oxford Instruments CT3500 cryostage. The images were
recorded with an electron dose of 10 e Å–2 and with a defocus range of 1–2 µm
at 30,000× on Kodak SO163 film. Images were digitized on a Leafscan 45 linear
1132
CCD scanner (Ilford) with a step size of 20 µm, which corresponds to a pixel
size of 6.67 Å on the specimen. Particle selection was done with the MRC program Ximdisp23. Three-dimensional reconstruction from a data set of 1,500
particle images was done by projection matching with SPIDER24, using the EM
structure of wild-type GroEL as a starting model9. The resolution was
measured by Fourier shell correlation, using the 0.5 correlation criterion.
Negative-stain EM. A Heraeus oven was modified to allow the preparation of
negative-stain grids at a known fixed temperature. The door of the oven was
replaced with a perspex front with two 15-cm circular holes to allow hand access.
A thermostatic oven controller (RS Components) and fan were added. In addition to the controller, temperature was monitored by a thermocouple in the
immediate vicinity of the sample and a mercury thermometer at the top of the
oven. The temperature of the chamber was 40 ± 2 °C during the experiment.
Stain and pipette tips were allowed 5 min to reach the set temperature. Grids were
prepared in the oven by depositing a 3-ml droplet onto the grid and then rinsing
with 3% w/v uranyl acetate. Images were recorded at 42,000× and 100 kV on a
Tecnai T10 electron microscope. Images were collected from two repeats of a time
course experiment over 2.5 h, comparing complexes formed with wild-type or
E461K GroEL (0.1 µM), GroES (>0.2 µM) and ATP (4 mM). A total of
2,000–5,000 particles were counted, independently by two observers, from each
micrograph.
Energy calculations. A coordinate set for E461K was generated by substituting
a lysine side chain for Glu461 in the GroEL structure (PDB entry 1OEL)25.
Hydrogen atoms were added using the HBUILD facility of CHARMM26, and all
energy calculations were done with CHARMM27 using the CHARMM 22
force-field28, with a distance-dependent dielectric to approximate the effect of
solvent. The coordinate model was generated from a pair of subunits, one from
each ring, the interactions with the remaining subunits being inferred from
C7 symmetry. For each angular displacement from the native orientation, an
energy minimization was done as follows: atoms >8 Å from the equatorial
plane were ‘fixed,’ but allowed to interact with the remainder of the system. The
energy was then minimized by 40 steps of steepest descent and 1,000 steps of
conjugate gradient minimization, after which the energies and coordinates
were written out. Changing to a constant dielectric of 1 or 4 did not qualitatively alter the shape of the energy profiles.
Coordinates. The EM density map of GroEL E461K has been deposited in the
EM database (accession code EMD-1095).
ACKNOWLEDGMENTS
We thank R. Westlake, D. Houldershaw and S. Terrill for computing support,
W. Fenton and M. Karplus for advice and discussion, M. Jaffer and W. Williams
for help with particle counting, and The Wellcome Trust and the Howard Hughes
Medical Institute for financial support.
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Received 1 June; accepted 15 September 2004
Published online at http://www.nature.com/nsmb/
1. Sigler, P.B. et al. Structure and function in GroEL-mediated protein folding. Annu.
Rev. Biochem. 67, 581–608 (1998).
2. Brinker, A. et al. Dual function of protein confinement in chaperonin-assisted protein
folding. Cell 107, 223–233 (2001).
3. Saibil, H.R. & Ranson, N.A. The chaperonin folding machine. Trends Biochem. Sci.
27, 627–632 (2002).
4. Todd, M.J., Lorimer, G.H. & Thirumalai, D. Chaperonin-facilitated protein folding:
optimization of rate and yield by an iterative annealing mechanism. Proc. Natl. Acad.
Sci. USA 93, 4030–4035 (1996).
5. Rye, H.S. et al. GroEL-GroES cycling: ATP and nonnative polypeptide direct alternation of folding-active rings. Cell 97, 325–338 (1999).
6. Gray, T.E. & Fersht, A.R. Cooperativity in ATP hydrolysis by GroEL is increased by
GroES. FEBS Lett. 292, 254–258 (1991).
7. Yifrach, O. & Horovitz, A. Nested cooperativity in the ATPase activity of the oligomeric
chaperonin GroEL. Biochemistry 34, 5303–5308 (1995).
8. Burston, S.G., Ranson, N.A. & Clarke, A.R. The origins and consequences of asymmetry in the chaperonin reaction cycle. J. Mol. Biol. 249, 138–152 (1995).
9. Roseman, A.M., Chen, S., White, H., Braig, K. & Saibil, H.R. The chaperonin ATPase
cycle: mechanism of allosteric switching and movements of substrate-binding
domains in GroEL. Cell 87, 241–251 (1996).
VOLUME 11 NUMBER 11 NOVEMBER 2004 NATURE STRUCTURAL & MOLECULAR BIOLOGY
© 2004 Nature Publishing Group http://www.nature.com/nsmb
ARTICLES
10. Rye, H.S. et al. Distinct actions of cis and trans ATP within the double ring of the
chaperonin GroEL. Nature 388, 792–798 (1997).
11. Ranson, N.A. et al. ATP-bound states of GroEL captured by cryo-electron microscopy.
Cell 107, 869–879 (2001).
12. Ditzel, L. et al. Crystal structure of the thermosome, the archaeal chaperonin and
homolog of CCT. Cell 93, 125–138 (1998).
13. Horwich, A.L., Low, K.B., Fenton, W.A., Hirshfield, I.N. & Furtak, K. Folding in vivo of
bacterial cytoplasmic proteins: role of GroEL. Cell 74, 909–917 (1993).
14. Fenton, W. A., Kashi, Y., Furtak, K. & Horwich, A. L. Residues in chaperonin GroEL
required for polypeptide binding and release. Nature 371, 614–661 (1994).
15. Sot, B., Galan, A., Valpuesta, J.M., Bertrand, S. & Muga, A. Salt bridges at the inter-ring
interface regulate the thermostat of GroEL. J. Biol. Chem. 277, 34024–34029 (2002).
16. Yifrach, O. & Horovitz, A. Two lines of allosteric communication in the oligomeric
chaperonin GroEL are revealed by the single mutation Arg196→Ala. J. Mol. Biol.
243, 397–401 (1994).
17. Weissman, J.S. et al. Mechanism of GroEL action: productive release of polypeptide
from a sequestered position under GroES. Cell 83, 577–587 (1995).
18. Ma, J., Sigler, P.B., Xu, Z. & Karplus, M. A dynamic model for the allosteric mechanism of GroEL. J. Mol. Biol. 302, 303–313 (2000).
19. Kafri, G. & Horovitz, A. Transient kinetic analysis of ATP-induced allosteric transitions
in the eukaryotic chaperonin containing TCP-1. J. Mol. Biol. 326, 981–987 (2003).
20. Braig, K. et al. The crystal structure of the bacterial chaperonin GroEL at 2.8 Å.
Nature 371, 578–586 (1994).
21. Horovitz, A., Bochkareva, E.S., Kovalenko, O. & Girshovich, A.S. Mutation Ala2→Ser
destabilizes intersubunit interactions in the molecular chaperone GroEL. J. Mol. Biol.
231, 58–64 (1993).
22. Ranson, N.A., Dunster, N.J., Burston, S.G. & Clarke, A.R. Chaperonins can catalyse
the reversal of early aggregation steps when a protein misfolds. J. Mol. Biol. 250,
581–586 (1995).
23. Crowther, R.A., Henderson, R. & Smith, J.M. MRC image processing programs.
J. Struct. Biol. 116, 9–16 (1996).
24. Frank, J. et al. SPIDER and WEB: processing and visualization of images in 3D electron microscopy and related fields. J. Struct. Biol. 116, 190–199 (1996).
25. Braig, K., Adams, P. D. & Brunger, A. T. Conformational variability in the refined
structure of the chaperonin GroEL at 2.8 Å resolution. Nat. Struct. Biol. 2,
1083–1094 (1995).
26. Brunger, A.T. & Karplus, M. Polar hydrogen positions in proteins—empirical energy
placement and neutron diffraction comparison. Proteins 4, 148–156 (1988).
27. Brooks B.R. et al. CHARMM: a program for macromolecular energy, minimization,
and dynamics calculations. J. Comp. Chem. 4, 187–217 (1983).
28. MacKerell, A.D. Jr. et al. All-atom empirical potential for molecular modeling and
dynamics studies of proteins. J. Phys. Chem. B 102, 3586–3616 (1998).
NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 11 NUMBER 11 NOVEMBER 2004
1133